#airship (2018-11)

@maarten has joined the channel

@Erik Osterman (Cloud Posse) has joined the channel

@jamie has joined the channel

@jonboulle has joined the channel

@Andriy Knysh (Cloud Posse) has joined the channel

Hi Everyone, welcome to the channel, thanks @Erik Osterman (Cloud Posse) ! Happy to help out Airship module users and to also discuss upcoming changes to the module here.

And happy to announce that the latest ECS Service 0.8.0 supports Cognito authentication, http to https redirection. and it can deal with ever changing task definitions by any external pipeline and still allows Terraform to modify the Environment Variables and/or cpu and mem.

@Nikola Velkovski has joined the channel

That’s really cool! Like the cognito integration. Should make it easier to expose staging apps

Exactly and drop nginx services with the sole purpose of providing a simple htaccess

Need some ideas here:

https://github.com/blinkist/terraform-aws-airship-ecs-service/blob/966757509ca42a0b4d3af541d7b5f864e8ff8ac4/modules/live_task_lookup/output.tf#L33

So I’ve added a lambda for live task lookup to overcome the bootstrapping issue. But it seems that a lambda_invocation datasource works differently from a regular datasource.

We create a Taskdefinition with an image coming from the live_lookup module, when using a normal datasource, it will figure out that when the taskdefinition is the same as the current one and will not create a new one. But when using the lambda invocation as datasource for getting the current image, the plan already describes it will create a new task_definition no matter what. UGH!

pfew, found the problem

I had a depends_on in the datasource to the lambda_function..

depends_on in datasources always force recreation..

so it all works now?

yeah

Do you think you can get Anatoli to take a look at the nodejs, if not I’ll have to find someone else

also, currently i’m creating the iam resources for the lambda in modules/iam, shall we leave it there or move it to the live_task_lookup

Maybe you can take another look at the logic itself in js

PR is out Jon, sorry that you need to go through Javascript code .. I’ve asked in the node subreddit and one guy with unknown seniority said the code was fine .

Fellas! I’m adding Scheduled jobs to the module, callable with simple list block of multiple jobs. Cloudwatch event triggers an event which hits a lambda, lambda figures out what the current task definition of the service is and will run the command on there. Let me know if you have ideas or concerns.

I think I leave out the container env vars override as it’s too much hassle now, same with placement constraints, as they can be added later.

I’m currently creating a different iam role for this lambda, but I’m opting for using one role defined in the iam module instead.

    ecs_scheduled_tasks = [{
      job_name                           = "vacuum"
      schedule_expression        = "cron(0 12 * * ? *)"
      command                           = "python vacuum_db.py"
      container_envvars_override = "A=B;VAR2=C"
    }]

Work in progress : https://github.com/blinkist/terraform-aws-airship-ecs-service/blob/ecs_scheduled_tasks/modules/ecs_scheduled_tasks/main.tf

0.8.1: Merge pull request #24 from blinkist/lambda_lookup Lambda lookup

hehe thanks @Erik Osterman (Cloud Posse)

funky

so every time someone clicks on the word variables.tf on slack

If I drop the redirect I can probably just banner it annoyingly

Hahjaha

Good one Martin

airship.tf

need to find out what slack gets and what not

Are you going to build a Jekyll site?

what’s that ?

“Transform your plain text into static websites and blogs. “

variables.tf is probably a good site for a blog

But for airship, a nice project site would be nice with some samples

https://help.github.com/articles/using-jekyll-as-a-static-site-generator-with-github-pages/

It’s the predominant way static sites for projects in GitHub are generated

Hundreds of templates out there

Do you have a favourite example of a project ?

It’s on my list of things to research. Plan to use our readme spec to generate static sites for all of our modules.

http://jekyllthemes.org/

nice, i’ll have a look.. First have to find a way to get the README.md a bit more attractive in general

There is also https://pages.github.com/themes/

funnily enough readme.md does not trick slack

Is there a .md TLD?

Looks like there is!

https://probot.github.io/ is a good one

yea, that one is nice

@jonboulle we now have merge commit set by default, shall we make it squash & merge ?

no strong opinion… I like merge commits for auditability and reverts but that only works with extremely strict commit diligence anyway

still no gif support here

0.8.2: Adding lambda_ecs_task_scheduler (#25) Init commit scheduled tasks

wip

ecs_scheduled_tasks update

small fixes

ecs_task_scheduler

ecs_scheduled_tasks clean up

Readme updates

fmt

Fix after PR Review

yeah this can be better

After going through tons of jekyll examples, I think https://vuepress.vuejs.org which is what is powering https://www.runatlantis.io looks to be the least time intensive with the most powerful layout there.

if considering just static site generation, did you consider Hugo?

https://github.com/gohugoio/hugo

main reason to use Jekyll is it’s supported by GitHub without any CI/CD systems

https://themes.gohugo.io/

(30K stars - very popular)

https://themes.gohugo.io/theme/hugo-fresh/

yeah but CICD is really the least of our problems, so it’s not a decision maker. I was looking for a simple DOC Layout, with a side bar etc. and this looks I don’t need to spend much time on it which is exactly what I want : https://github.com/runatlantis/atlantis/tree/master/runatlantis.io

@jonboulle @jamie replied to your pr comments, if you have time to take a look .

Hi, thanks. Ill look later today.

https://aws.amazon.com/blogs/compute/introducing-aws-app-mesh-service-mesh-for-microservices-on-aws/

requires awsvpc

think we built that in if there now only would be instances with a bit more eni capability ..

@davidvasandani has joined the channel

Hi @davidvasandani !

@geertn has joined the channel

@jamie Do you have that initial setup of the service registry again for me, slack history doesn’t show it anymore. Also in your original PR there are some things volume related which I will cover after service discovery.

0.8.3: NLB Services (#26) Network Load Balancing Support added for the ECS Module. The parameter load_balancing_type can be set to either “none”, “network” or “application” and takes care of attaching the related LB type.

@jamie and another question, for a service to connect to other services in the service registry, it does not need to be registered to the registry itself, correct ?

correct